International Journal of Colorectal Disease (2019) 34:451–457 https://doi.org/10.1007/s00384-018-3212-6

ORIGINAL ARTICLE

Surgical outcomes of patients treated with vs. vedolizumab in inflammatory bowel disease: a matched case analysis

Matteo Novello1 & Luca Stocchi1 & Stefan Holubar1 & Sherief Shawki1 & Jeremy Lipman1 & Emre Gorgun1 & Tracy Hull1 & Scott R. Steele1

Accepted: 30 November 2018 /Published online: 10 December 2018 # Springer-Verlag GmbH Germany, part of Springer Nature 2018

Abstract Aim The association between preoperative use of monoclonal in inflammatory bowel disease (IBD) patients and postoperative complications is controversial, especially for the latest approved biologics, ustekinumab and vedolizumab, where data is limited. We hypothesized that ustekinumab-treated patients would have a similar overall postoperative complication rate as vedolizumab-treated patients. The aim of this study was to compare postoperative complications in patients receiving preop- erative ustekinumab vs. vedolizumab. Methods We queried our IRB-approved prospective database to identify Crohn’s patients who underwent colorectal surgery and pretreatment with ustekinumab or vedolizumab within 12 weeks of surgery. Ustekinumab-treated patients were matched to vedolizumab-treated patients based on sex, age ± 5 years, date of operation ± 3 years, and type of surgery. Paired univariate analysis and conditional logistic regression of the matched pairs were performed. Our primary outcome was the short-term postoperative complication rate. Secondary outcomes included infectious complications, readmission, reoperation, and length of stay. Results A total of 103 patients with Crohn’s disease (CD) met the inclusion criteria (mean age 38.7 ± 13.4 years; male 51.2%). Overall, 30 patients received preoperative ustekinumab and 73 vedolizumab. In the univariate analysis, vedolizumab-treated patients had a higher postoperative complication rate (p = 0.009) and ileus rate (p = 0.015). After matching, 26 matched pairs were compared and logistic regression models demonstrated no significant difference in the primary outcome. Conclusions For our limited experience, the choice of preoperative biologic treatment between ustekinumab and vedolizumab should not be influenced by fear of surgical complications.

It was presented as an E-poster at BThe American Society of Colon and Rectal Surgeons 2018 Annual Scientific Meeting^ in Nashville, TN, USA, May 19–23, 2018. What does this paper add to the literature? The association between preoperative use of ustekinumab and vedolizumab in inflammatory bowel disease patients and postoperative complications is controversial. Considering the little data available in literature and the increasing use of these biologics, we believe it is important to shed light on their preoperative use with respect to postoperative complications.

* Scott R. Steele Jeremy Lipman [email protected] [email protected] Matteo Novello Emre Gorgun [email protected] [email protected] Luca Stocchi [email protected] Tracy Hull [email protected] Stefan Holubar [email protected] 1 Department of Colorectal Surgery, Digestive Disease & Surgery Sherief Shawki Institute, Cleveland Clinic, 9500 Euclid Ave, Cleveland, OH 44195, [email protected] USA 452 Int J Colorectal Dis (2019) 34:451–457

Keywords Ustekinumab . Vedolizumab . Biologics . Colorectal surgery . Outcomes . Morbidity . Mortality . Comparison

Introduction that USTE-treated patients would have a similar overall post- operative complication rate as compared to VEDO-treated One of the most common indications of surgical intervention in patients. IBD is refractoriness to medical treatment. Many biologic agents have been introduced sequentially to the medical arma- mentarium in managing (UC) and Crohn’s Methods disease (CD). Available in 2014, vedolizumab (VEDO) is a human monoclonal that selectively inhibits leukocyte We conducted a retrospective, single tertiary center, case- adhesion and migration into the gastrointestinal tract by bind- matched study of a prospectively maintained database. We ingtotheα4β7 . It was implemented in treating CD queried our IRB-approved prospective database to identify and UC as escalation therapy after failing other conventional patients with CD undergoing abdominal surgery from medical treatments. Subsequently, ustekinumab (USTE), a hu- May 2014 until October 2017, corroborated by chart review. man which binds to -12 and We stratified patients into two groups on the basis of mono- interleukin-23 preventing inflammatory cellular activation, was clonal antibody treatment within 12 weeks from surgery: Food and Drug Administration–approved in 2016, though it is VEDO vs. USTE. Patients undergoing anorectal surgery alone currently approved only for CD. and those who had treatment with anti-TNF-α agents within While these agents were shown to achieve remission in 12 weeks from surgery were excluded. The primary endpoint 53.1% [1] (USTE, CD) and 44.8% [2] (VEDO, CD) of pa- was overall morbidity occurring within the first 30 days after tients respectively, some patients will not respond and prog- colorectal surgery for IBD. Secondary outcomes included 30- ress to surgery. Usually, this happens in the context of refrac- day infectious complications, readmission, reoperation, and tory disease or progression while on medical therapy, indicat- postoperative length of stay. ing failure of medical management prompting surgical inter- Data abstracted included patient demographics, gender, age vention. In this event, surgery is usually offered while the at the time of surgery, smoking history, comorbidities, dura- patient is still receiving the biologic agent. While the type of tion of VEDO therapy, surgical procedure details, postopera- surgery needed depends on the disease and its sequelae, tive complications, postoperative packed red blood cell whether it is UC or CD, the impact of the preoperative bio- (PRBC) transfusion, postoperative length of hospital stay, re- logic on the perioperative course and complications remains operation, readmission, and mortality. Specific postoperative controversial. complications occurring within 30 days after surgery included After performing a study at our department, we noticed that surgical site infections (SSI, superficial or organ space), fas- preoperative VEDO is comparable to other biologics, such as cial dehiscence, perineal wound infection, anastomotic leak, and , with respect to the postoperative other enteric leak, rectal stump leak, pneumonia, Clostridium complication rate. However, there are two other similar stud- difficile infection, sepsis, sterile fluid collection, stoma com- ies available in the literature regarding VEDO which present plication, small-bowel obstruction (SBO), ileus, bleeding, de- partially contradictory conclusions [3, 4]. Lightner and col- hydration, and urinary tract infection (UTI). The variable leagues reported the link between the use of preoperative BAny infection^ was defined as the patient having at least VEDO and risk of surgical site infection (SSI) in patients with one of the following considering the Center of Disease IBD, while Yamada and colleagues stated that preoperative Control definitions [6]: superficial SSI and/or organ space VEDO administration did not influence the postoperative SSI (intra-abdominal and/or pelvic abscess). Ileus was defined morbidity rate. Additionally, the first study evaluating preop- as the absence of bowel function on postoperative day 5 or the erative USTE was recently published and did not find a sig- need for nasogastric tube insertion due to abdominal disten- nificant difference in terms of SSI when comparing it to pre- sion, nausea, or vomiting after initiating a liquid diet without operative anti-TNF-α exposure [5]. The administration of this evidence of mechanical bowel obstruction [7]. Dehydration new generation of biologics is increasing, and it is necessary was defined as ostomy output of 1500 mL over 24 h and a to optimize the preoperative treatment of this patient popula- blood urea nitrogen/creatinine ratio ≥ 20 and/or physical find- tion. To our knowledge, no study had compared the pattern of ings of dehydration [8]. Hospital readmissions and postop complications associated with preoperative treatment reoperations were counted when occurring within 30 days with USTE to VEDO. from the date of hospital discharge. We collected data on pre- Therefore, the aim of this study was to compare postoper- operative medication use within 1 month from surgery for ative complications after colorectal surgical procedures for corticosteroids, opioid analgesics, and mesalamine (and deriv- IBD in patients receiving VEDO vs. USTE. We hypothesized atives). Preoperative immunomodulator use (/6- Int J Colorectal Dis (2019) 34:451–457 453 mercaptopurine, ) was considered positive when previously treated with VEDO; thus, the number of other bi- within 2 months from surgery. In addition, current or past use ologics failed before surgery was significantly different be- of biologic agents was carefully reported. The variable tween the cohorts (p < 0.001). Regarding differences in base- BBiologics tried any time before surgery^ was not recorded line and operative characteristics, the VEDO-treated patients for the vedolizumab group. Nutritional status was evaluated presented more frequently with BAnemia/bleeding/preopera- by collecting serum albumin within 30 days from the tive transfusion required^ (15.8% vs. 0%, p = 0.0022). The operation. most common category of procedure was group 1 (N = 35, Case matching was performed to directly match USTE- 48.6%). Baseline characteristics after matching are shown in treated patients to VEDO-treated patients on a 1:1 basis. The Table 2. Overall, 26 matched pairs were successfully con- matching criteria were age ± 5 years, surgery date ± 2 years, structed and there was no difference in age, surgery date, gen- gender, and procedure. The procedures selected were all elec- der, diagnosis, and procedures. Preoperative use of immuno- tive and the three groups were defined as follows: group 1, modulators, corticosteroids, opioids, mesalamine, completion proctectomy (CP) + end ileostomy (EI), total ab- albuminemia, and comorbidities was all comparable among dominal colectomy (TAC) + EI, proctectomy + EI, total the groups in the matched paired univariate analysis. There proctocolectomy (TP) + EI, Hartmann’s procedure, ostomy was no significant difference among the matched pairs when procedure, TAC with ileorectal anastomosis + ostomy; group comparing the number of patients who had other biologics 2, segmental resection, ileocolic resection, enterectomy; group than those who had USTE/VEDO respectively within 3, strictureplasty isolated; group 4, other pouch procedures. 12 weeks (p = 0.66) and 1 year from surgery (p =0.56). We selected these combinations so that in group 1, we could With respect to the unmatched outcomes (Table 3), VEDO- include major resections with stoma and in group 2 segmental treated patients had a higher overall postoperative complica- resections. Group 4 was not densely populated and was ex- tion rate (54.8% vs. 26.7%, p = 0.009) and an increased ileus cluded from the matching process because of no matching rate (23.3% vs. 3.3%, p = 0.015). A total of 4 (5.5%) VEDO- patients available in the VEDO group. treated patients required reoperation compared with zero in the USTE group (p =0.59). Statistical analysis A linear mixed model showed that patients treated with USTE stayed 1.1 days less than patients who received Univariate analyses were conducted to compare the demo- VEDO; however, this was not statistically significant (95% graphics and preoperative variables of the USTE and VEDO confidence interval [CI] − 4.4, 2.3; p = 0.52). Eleven percent groups. Pearson’s chi-square test or Fisher’s exact test was (N = 3) of the VEDO-treated patients vs. 59% (N = 16) of the used for categorical factors, and ANOVA or the Kruskal- USTE-treated patients had an open procedure performed (p = Wallis test was applied for continuous factors as appropriate. 0.002): ostomy procedures were more common in the USTE A descriptive paired univariate analysis was conducted on the group (30.1%, N = 8) than in the VEDO group (11.1%, N = 3), matched sample to analyze the balance of other patient char- even though this difference was not statistically significant. In acteristics between treatment groups. Paired Student’s t test or addition, the increased ileus rate identified in the univariate Wilcoxon signed-rank test was used for continuous factors, unmatched analysis was not confirmed at the paired univariate and McNemar’stestsorBowker’s test of symmetry was used analysis (p =0.21). for categorical variables. For the paired analysis of outcomes, the conditional logistic A conditional logistic regression model of the matched regression model (Table 4) showed no significant difference pairs was fit to assess readmission, reoperation, and overall among the matched pairs for overall complications, infectious and specific complications. Lastly, a linear mixed model complication, readmission rate, dehydration, enteric leak, treating the matched pairs as a random effect was fit to eval- intra-abdominal abscess, perineal wound infection, pneumo- uate the impact of treatment on length of stay. p values < 0.05 nia, sepsis, organ space SSI, and superficial SSI. were considered statistically significant. All analyses were performed using SAS (version 9.4, The SAS Institute, Cary, NC). Discussion and conclusion

This study demonstrated that IBD cases selected to undergo Results surgery following treatment with USTE had a similar risk of overall postoperative complications and perioperative out- A total of 103 patients with CD were included (mean age 38.7 comes as compared to VEDO administration. We designed ± 13.4 years; male 46.6%). Overall, 73 were treated with our study to evaluate the introduction into our practice of VEDO preoperatively and 30 received USTE (Table 1). two new biologic medications with very different pharmaco- Thirty-seven percent of cases in the USTE group were dynamics from the most commonly used anti-TNF-α agents, 454 Int J Colorectal Dis (2019) 34:451–457

Table 1 Demographics and preoperative variables of Factor VEDO (N =73) USTE(N =30) p value unmatched retrospective cohorts Male gender 36 (49.3) 12 (40.0) 0.39c Age at surgery (years) 39.7 ± 13.3 36.3 ± 13.7 0.25a Body mass index (kg/m2) 24.9 ± 6.2 23.4 ± 7.1 0.29a Race 0.32d Hispanic 2 (2.7) 2 (6.7) Not Hispanic 70 (95.9) 28 (93.3) Duration of therapy before surgery (days) 208.6 ± 186.9 NA NA Time from last dose to surgery (days) 41.1 ± 23.8 50.9 ± 30.1 0.082a Pulmonary comorbidity 4 (5) 0 (0) 0.35c Cardiovascular comorbidity 8 (10.7) 4 (12.5) 0.75c Renal comorbidity 1 (1.4) 0 (0) 1.0c Anemia/bleeding/preoperative transfusion required 9 (15.8) 0 (0) 0.022c Preoperative albumin (g/dL) 3.6 ± 0.6 3.7 ± 0.4 0.39a Diabetes 3 (4.1) 0 (0) 0.99d Number of other biologics failed before surgery 2.0 (0.00, 4.0) 3.0 (0.00, 4.0) < 0.001b Biologics tried any time before surgery NA Adalimumab NA 27 (90.) Infliximab NA 26 (86.7) Vedolizumab NA 11 (36.7) Certolizumab NA 12 (40) Procedure group 0.54d Group 1 35 (48.6) 14 (46.7) Group 2 34 (47.2) 13 (43.3) Group 3 2 (2.8) 1 (3.3) Group 4 1 (1.4) 2 (6.7) Any new stoma 59 (80.8) 19 (63.) 0.07c Other biologics within 3 months from surgery 3 (4.1) 4 (13.3) 0.66c Other biologics within 1 year from surgery 33 (45.2) 15 (50.0) 0.70c Preoperative immunomodulators 18 (24.7) 9 (30.0) 0.58c Preoperative mesalamine (and derivatives) 13 (17.8) 2 (6.7) 0.15c Preoperative corticosteroids 37 (50.7) 14 (46.7) 0.71c Opioids within 1 month preoperatively 28 (38.4) 11 (36.7) 0.87c

Statistics are presented as mean ± SD, median (min, max), or N (column %) p values: a = ANOVA, b = Kruskal-Wallis test, c = Pearson’s chi-square test, d = Fisher’s exact test. P <0.05 but also concerning for leading to perioperative complications based on previous reports from the Mayo Clinic [3, 5]. We [3, 5]. VEDO was selected as the control group medication believe that the time difference between the two half-lives is since it presents similar indications in the management of the not large enough to justify a clinically significant influence on patient with refractory CD with respect to USTE. In addition, the postoperative outcomes. they are both representatives of a new generation of non-anti- The case-matched approach allowed us to carefully balance TNF biologics. When comparing the effects of biologic the two cohorts and compare patients with similar character- agents, it is important to take into account the half-life of the istics. We case-matched based on demographics, date of sur- antibody since it might have a substantial impact on the pa- gery, and specific procedures performed. We excluded indi- tient’s conditions. The medications must be both taken every viduals who underwent surgery without receiving biologics 8 weeks, and their half-life is similar, being approximately and those who had treatment with anti-TNF-α agents within 19 days for USTE [9] and 25 days for VEDO [10]. 12 weeks from surgery since this would have triggered a fur- Additionally, the duration from last dose to the subsequent ther selection bias in an already diverse patient population. In surgery date was comparable between our study groups and fact, our results show that USTE and VEDO are typically used never exceeded 12 weeks, since this was an inclusion criterion as fourth- or fifth-line biologic agents at our institution, and Int J Colorectal Dis (2019) 34:451–457 455

Table 2 Demographics and preoperative variables of matched Factor VEDO (N = 26) USTE (N =26) p value retrospective cohorts Male gender 11 (42.3) 11(42.3) – Age at surgery (years) 38.5 ± 11.7 38.4 ± 13.2 0.96a Body mass index (kg/m2) 23.6 ± 5.6 23.7 ± 7.6 0.95a Race 0.56c Hispanic 1 (3.8) 2 (7.7) Not Hispanic 25 (96.2) 24 (92.3) Time from last dose to surgery (days) 40.3 ± 23.1 49.5 ± 30.1 0.29a Number of other biologics failed before surgery 1.00 (0.00, 3.0) 3.0 (1.00, 4.0) < 0.001b Preoperative albumin (g/dL) 3.5 ± 0.7 3.7 ± 0.5 0.28a Diabetes 2 (7.7) 0 (0.0) 0.24c Biologics tried any time before surgery NA Adalimumab NA 24 (92.3) Infliximab NA 23 (88.5) Vedolizumab NA 11 (42.3) Certolizumab NA 12 (46.2) Procedure group – Group 1 14 (53.8) 14 (53.8) Group 2 11 (42.3) 11 (42.3) Group 3 1 (3.8) 1 (3.8) Any new stoma 24 (84.6) 17 (65.4) 0.13d Other biologics within 3 months from surgery 3 (11.5) 4 (15.4) 0.66c Other biologics within 1 year from surgery 16 (61.5) 14 (53.8) 0.56c Preoperative immunomodulators 6 (23.1) 8 (30.8) 0.48c Preoperative mesalamine (and derivatives) 3 (11.5) 2 (7.7) 0.65c Preoperative corticosteroids 9 (34.6) 11 (42.3) 0.53c Opioids within 1 month preoperatively 12 (46.2) 11 (42.3) 0.78c

Statistics are presented as mean ± SD, median (min, max), or N (column %) p values: a = Paired t test, b = Wilcoxon signed-rank test, c = McNemar’s test, d = Bowker’s test of symmetry. P <0.05 most of the patients received either medication after failing [5]. Lightner and colleagues concluded that USTE treat- treatment with anti-TNF-α agents. The data that 40% of the ment within 12 weeks of surgery did not appear to increase USTE-treated patients were previously treated with VEDO the risk of infectious complications above that of CD pa- frames USTE as a last therapeutic resource given the novelty tients treated with anti-TNF-α medications. In the univar- that it represents and the consequent lack of data in the liter- iate analysis, the USTE-treated patients had a significantly ature. Moreover, this fact is supported by the remarkable rates higher rate of return to the operating room; however, none of preoperative administration of narcotics (N = 55, 32.0%) of the examined variables were significant in the multivar- and corticosteroids (N = 105, 61.0%) which, besides suggest- iate analysis. As stated by the authors, that study had sev- ing a generalized advanced disease status, have also been eral limitations. Data were obtained from six institutions, shown to increase postoperative infections rates [11–16]. which might imply a variability of the overall patient man- Our department showed in a case-matched study, not yet agement. It is striking that patients were enrolled during published, that preoperative VEDO is comparable to other clinical trials, before the final FDA approval of USTE, biologics, such as infliximab and adalimumab, with respect thereby possibly assuming different doses/regimens than to the postoperative complication rate. Thus, such results the officially approved ones [1]. In addition, the operations are closer to those reported by Yamada [4] and colleagues performed were grouped under the definition of Bmajor who demonstrated that preoperative VEDO administration abdominal procedure^ which is rather non-specific and did not influence the postoperative morbidity rate of pa- might as well refer to non-colorectal procedures. The spe- tients undergoing surgery for IBD. To the best of our cific surgery performed may be significantly associated knowledge, there is only one study in the literature regard- with the risk of postoperative complications [17]. We ing the safety of preoperative USTE in colorectal surgery therefore tried to avoid this confounding factor by 456 Int J Colorectal Dis (2019) 34:451–457

Table 3 Overall postoperative outcomes Variable VEDO (N = 73) USTE (N =30) p value

Length of stay 8.4 ± 6.6 6.9 ± 5.3 0.27a Any postoperative complication 40 (54.8) 8 (26.7) 0.009c Overall infectious complications 11 (15.1) 7 (23.3) 0.32c Sepsis 3 (4.1) 1 (3.3) 0.99d Organ space SSI 4 (5.5) 1 (3.3) 0.99d Superficial SSI 6 (8.2) 3 (10.0) 0.77c Readmission 10 (13.7) 2 (6.7) 0.31c Reoperation 4 (5.5) 0(0.0) 0.59d

Statistics are presented as mean ± SD, median [P25, P75], or N (column %) p values: a = ANOVA, b = Kruskal-Wallis test, c = Pearson’s chi-square test, d = Fisher’s exact test. P <0.05 enrolling only cases of elective abdominal colorectal pro- cause, and effect association and further evaluation with lon- cedures, which were case-matched according to the type of ger follow-up and larger experience will determine if this re- procedure performed. mains the same. Our study was limited by its single institution non- Despite these limitations, this is the first study in the liter- randomized retrospective nature. It is important to point out ature comparing USTE and VEDO from a surgical point of that VEDO is FDA-approved for both UC and CD, while view. Considering the short study time frame, we were able to USTE is only approved for CD. This indication might expand provide a first insight on these cases only thanks to the high to include UC largely depending on the results of the UNIFI volume and complexity level of cases managed at our unit. trial, the results of which should be available in November Based on the findings of this study, our department continues 2021 [18]. Nevertheless, we included only Crohn’spatients to look closely at results but considers surgery with a stoma and tried to address the heterogeneity of our patient sample by not always mandatory. However, larger and more powerful performing a case-matched analysis. As a tertiary care facility study designs are necessary to generalize the results and clar- that has a more complex referral base, our results may not ify the relationship between surgical outcomes and the admin- represent other institutions as a whole. Further, there is a small istration of this new generation of biologics. possibility that some of these patients may ultimately carry a In conclusion, our data indicate that the preoperative use of diagnosis of indeterminate colitis and not CD. Since USTE USTE in IBD patients, within 12 weeks from surgery, is as- was approved by the FDA in September 2016, our project sociated with similar perioperative outcomes and overall post- includes just 12 months of data, thereby limiting the statistical operative complications when compared to VEDO. Thus, for power of the model. Our study may be underpowered to detect our limited experience, the choice of preoperative biologic a difference or rather a failure to identify a true difference as treatment between USTE and VEDO should not be influenced seen in the univariate analysis, and there is the potential risk of by fear of surgical complications. bias given the heterogeneity of the grouped cohorts. In addition, there was an association with an increased Contribution of each author/coauthor number of patients with anemia or bleeding in the VEDO a. Substantial contributions to the conception or design of group. At the present time, the authors feel this is not a true the work or the acquisition, analysis, or interpretation of data for the work: Novello M., Stocchi L., Holubar S.D., Table 4 Postoperative outcomes in the case-matched (1:1) logistic Shawki S., Lipman J.M., Gorgun E., Hull T., Steele S.R. regression for ustekinumab vs. vedolizumab b. Drafting the article or revising it critically for important intellectual content: Novello M., Stocchi L., Holubar S.D., Variable Odds ratio (95% CI) p value Shawki S., Lipman J.M., Gorgun E., Hull T., Steele S.R. Any postoperative complication 0.38 (0.10, 1.4) 0.15 c. Final approval of the version to be published: Novello M., Overall infectious complications 1.5 (0.42, 5.3) 0.53 Stocchi L., Holubar S.D., Shawki S., Lipman J.M., Sepsis 1.00 (0.06, 16.0) 0.99 Gorgun E., Hull T., Steele S.R. d. Organ space SSI 0.50 (0.05, 5.5) 0.57 Agreement to be accountable for all aspects of the work in Superficial SSI 1.00 (0.14, 7.1) 0.99 ensuring that questions related to the accuracy or integrity Readmission 0.25 (0.03, 2.2) 0.21 of any part of the work are appropriately investigated and resolved: Novello M., Stocchi L., Holubar S.D., Shawki N = 27 patients in each group S., Lipman J.M., Gorgun E., Hull T., Steele S.R. Int J Colorectal Dis (2019) 34:451–457 457

Compliance with ethical standards Dec 28];21(8):641–8. Available from: https://doi.org/10.1007/ s10151-017-1667-z Conflict of interest The authors declare that they have no conflict of 9. Food and Drug Administration. STELARA® (ustekinumab) injec- interest. tion, for subcutaneous or intravenous - FDA label [Internet]. 2016 [cited 2018 Apr 7]. Available from: https://www.accessdata.fda. gov/drugsatfda_docs/label/2016/761044lbl.pdf 10. Food and Drug Administration, Center for Drug Evaluation and References Research. ENTYVIO (vedolizumab) for injection, for intravenous use - FDA label [Internet]. 2017 [cited 2018 Apr 7]. Available from: 1. Feagan BG, Sandborn WJ, Gasink C, Jacobstein D, Lang Y, https://www.accessdata.fda.gov/drugsatfda_docs/label/2014/ Friedman JR, et al. (2016) Ustekinumab as induction and mainte- 125476s000lbl.pdf nance therapy for Crohn’s disease. N Engl J Med [Internet]. [cited 11. Li Y, Stocchi L, Cherla D, Liu X, Remzi FH (2016) Association of 2018 Jan 21];375(20):1946–60. Available from: https://doi.org/10. preoperative narcotic use with postoperative complications and 1056/NEJMoa1602773 prolonged length of hospital stay in patients with Crohn disease. 2. Sandborn WJ, Feagan BG, Rutgeerts P, Hanauer S, Colombel J-F, JAMA Surg [Internet]. [cited 2018 Jan 21];151(8):726–34. Sands BE, et al. (2013) Vedolizumab as induction and maintenance Available from: http://www.ncbi.nlm.nih.gov/pubmed/26913479 therapy for Crohn’s disease. N Engl J Med [Internet]. [cited 2018 12. Golub R, Golub RW, Cantu RJ, Stein HD (1997) A multivariate Apr 5];369(8):711–21. Available from: https://doi.org/10.1056/ analysis of factors contributing to leakage of intestinal anastomoses. NEJMoa1215739 J Am Coll Surg 184(4):364–372 3. Lightner AL, Raffals LE, Mathis KL, Cima RR, Tse CS, Pemberton 13. Post S, Betzler M, von Ditfurth B, Schürmann G, Küppers P, JH, et al. (2016) Postoperative outcomes in vedolizumab-treated Herfarth C (1991) Risks of intestinal anastomoses in Crohn’sdis- patients undergoing abdominal operations for inflammatory bowel ease. Ann Surg 213(1):37–42 disease. J Crohns Colitis [Internet]. [cited 2017 may 7];11(2):1–6. 14. Aberra FN, Lewis JD, Hass D, Rombeau JL, Osborne B, Available from: http://www.ncbi.nlm.nih.gov/pubmed/27543504 Lichtenstein GR (2003) Corticosteroids and immunomodulators: 4. Yamada A, Komaki Y, Patel N, Komaki F, Aelvoet AS, Tran AL, postoperative infectious complication risk in inflammatory bowel et al. (2017) Risk of postoperative complications among inflamma- disease patients. Gastroenterology 125(2):320–327 tory bowel disease patients treated preoperatively with 15. Yamamoto T, Allan RN, Keighley MRB (2000) Risk factors for vedolizumab. Am J Gastroenterol [Internet]. [cited 2018 intra-abdominal sepsis after surgery in Crohn’s disease. Dis Colon Jan 10];112(9):1423–9. Available from: https://doi.org/10.1038/ Rectum [Internet] 43(8):1141–1145 Available from: http://content. ajg.2017.201 wkhealth.com/linkback/openurl?sid=WKPTLP:landingpage&an= 5. Lightner AL, McKenna NP, Tse CS, Hyman N, Smith R, Ovsepyan 00003453-200043080-00017 G, et al. (2017) Postoperative outcomes in ustekinumab-treated 16. Nguyen GC, Elnahas A, Jackson TD (2014) The impact of preop- ’ patients undergoing abdominal operations for Crohn s disease. J erative steroid use on short-term outcomes following surgery for ’ Crohn s Colitis [Internet]. Available from: https://doi.org/10.1093/ inflammatory bowel disease. J Crohns Colitis [Internet] 8(12): ecco-jcc/jjx163/4694103 1661–1667 Available from: http://www.ncbi.nlm.nih.gov/pubmed/ 6. Centers for Disease Control and Prevention National Healthcare 25107847 Safety Network. Surgical site infection (SSI) event. [Internet]. 17. Varela JE, Wilson SE, Nguyen NT (2010) Laparoscopic surgery 2017 [cited 2017 Dec 27]. Available from: https://www.cdc.gov/ significantly reduces surgical-site infections compared with open nhsn/pdfs/pscmanual/9pscssicurrent.pdf surgery. Surg Endosc Other Interv Tech [Internet]. [cited 2016 7. Gu J, Stocchi L, Remzi FH, Kiran RP. (2014) Total abdominal Aug 7];24(2):270–6. Available from: http://www.ncbi.nlm.nih. colectomy for severe ulcerative colitis: does the laparoscopic ap- gov/pubmed/19533235 proach really have benefit? Surg Endosc [Internet]. [cited 2017 18. A study to evaluate the safety and efficacy of ustekinumab induc- Dec 26];28(2):617–25. Available from: https://doi.org/10.1007/ tion and maintenance therapy in participants with moderately to s00464-013-3218-7 severely active ulcerative colitis (UNIFI) - ClinicalTrials.gov 8. Li W, Stocchi L, Cherla D, Liu G, Agostinelli A, Delaney CP, et al. [Internet]. [cited 2018 Mar 11]. Available from: https:// (2017) Factors associated with hospital readmission following di- clinicaltrials.gov/ct2/show/study/NCT02407236 verting ileostomy creation. Tech Coloproctol [Internet]. [cited 2017